Context Engineering Toolkit

An interactive CLI and framework for setting up AI-ready projects with pre-configured MCP servers, Skills, and Agents.

🚀 Installation

Global Install (Recommended)

npm install -g context-engineering-installer

Run via NPX

npx context-engineering-installer

🛠️ Features

• Interactive CLI: Browse and install AI assets (Skills, Agents, Commands) directly into your project.
• MCP Management: Easily add and configure Model Context Protocol servers.
• Pre-configured Assets: A collection of "gold-standard" prompts and agents for common development tasks.
• Project Context: Tools to help AI assistants understand your project's unique patterns and rules.

📂 Structure

.github/
  skills/       # Reusable AI skill prompts
  agents/       # Autonomous AI agent definitions
  commands/     # Slash commands for quick actions
.vscode/
  mcp.json      # MCP Server configuration
cli/            # The interactive TUI application
AGENTS.md       # Core rules for AI assistants
EXTENDING.md    # Guide for adding new assets
README.md       # This file

🤖 Usage

1. Run the CLI: context-cli (if installed globally) or npx context-engineering-installer.
2. Browse Assets: Use the interactive TUI to explore available skills and agents.
3. Install: Select an asset to install it into your current project.
4. Configure MCP: Add servers like Context7 or GitHub to your .vscode/mcp.json.

OpenCode MCP servers

• This repository includes an opencode.json listing local and remote MCP servers (mirrored from .vscode/mcp.json) so OpenCode can discover the same tools. You can enable/disable servers there and provide any required environment variables (for example PERPLEXITY_API_KEY).

VS Code Copilot

• For VS Code Copilot compatibility the project provides recommended settings in .vscode/settings.json.

Images

➕ Adding New Assets

Want to add your own Skills, Agents, or MCP Servers? See EXTENDING.md for a complete guide.

📝 License

MIT